Re: [sqlite] sqlite_stat4 table

On 10/29/17, korablev wrote: > > Probably the main question is not "why there are 2 samples for 1 value", but > "why there is so strange (if not wrong) statistics?". > STAT4 is designed to help SQLite run faster for queries of tables with millions of rows. For a table

Re: [sqlite] sqlite_stat4 table

Richard Hipp-3 wrote > See https://www.sqlite.org/fileformat.html#the_sqlite_stat4_table Thanks, I have already read it, but there is no explanation about this: korablev wrote > statistics in the second row seems to be wrong: there are 0 rows which are > less than 1(for the reason that the only

Re: [sqlite] sqlite_stat4 table

On 10/29/17, korablev wrote: > DROP TABLE IF EXISTS t1; > create table t1(id primary key, a, b) without rowid; > create index t1a on t1(a); > insert into t1 values(1, 1, 2), (2, 1, 3); > analyze t1; > select * from sqlite_stat4 where idx = 't1a'; > > t1|t1a|2 1|0 0|0 0|

[sqlite] sqlite_stat4 table

DROP TABLE IF EXISTS t1; create table t1(id primary key, a, b) without rowid; create index t1a on t1(a); insert into t1 values(1, 1, 2), (2, 1, 3); analyze t1; select * from sqlite_stat4 where idx = 't1a'; t1|t1a|2 1|0 0|0 0| t1|t1a|2 1|0 1|0 1| Firstly, there is no blob values for